[Purpose/Significance] The discipline-oriented in-depth aggregation and navigation service of network information resources is an important issue in the field of Library and Information Science.This paper reviews the book"Discipline-oriented Research on In-depth Aggregation and Service of Network Information Resources", which aims to provide a reference for the research and practice in related fields under the new environment. [Method/Process] The core ideas and viewpoints of the book were reviewed from the new proposed research perspectives, technologies, methods, and applications. [Result/Conclusion] The publication of the book plays an important guiding role in the aggregation of subject network information resources and the construction of subject navigation.

[Purpose/Significance] Collaborative filtering recommendation based on nearest users, one of the most widely used algorithms in recommender systems, is affected by the issues of data sparsity and computational scalability, and the recommendation effect is unsatisfactory. [Method/Process] To address these issues, a category preferred data field clustering based collaborative filtering recommendation algorithm(CPDFC-CFR)was proposed.First, the algorithm discarded user ratings and used comment sentiment to construct a user-item matrix to enhance the ability to express user preferences.Second, the algorithm introduced the concepts of category preference and semantic preference, reduced the dimensionality of the user-item matrix using category preferred ratio, and incorporated comment sentiment preference, category preference, and semantic preference in the user similarity calculation to reduce data sparsity.Finally, the algorithm used the data field as the pre-algorithm for user clustering and used its output(maximum point)as an input to the K-means algorithm to improve the real-time and stability performance of the algorithm. [Result/Conclusion] The findings indicated that: 1the lower the item category level, the higher the accuracy(F-measure)and computational efficiency(number of similarity calculations and time-consuming of recommendation)of the CPDFC-CFR algorithm; 2compared with other recommendation algorithms, CPDFC-CFR algorithm can effectively improve the recommendation accuracy and computational efficiency, which is an important reference value for the construction of collaborative filtering recommendation system.

[Purpose/Significance] In computational chemistry, chemical bond energy is an important scientific data in chemistry.At present, the extraction of chemical bond energy is mainly carried out manually by experts with low efficiency.Most of the chemical bond energy data are hidden in numerous literatures, which brings difficulties for in-depth and innovative scientific data analysis. [Method/Process] In order to solve this problem, this study took the ChemBE chemical bond energy corpus as experiment dataset, and designed a transfer learning method to automatically extract the chemical bond energy data with little manual effort of domain expert.This paper put forward an end-to-end BERT-CRF model where domain high-frequency sub-words were collected to solve the problem of the unknown words from a large number of academic articles.In the subsequent training of the proposed model, the field characteristics of those sub-words were input into the model.Therefore, the implementation of the model can automatically and efficiently extract the chemical bond energy from academic articles. [Result/Conclusion] Comparing with baseline models, experiment results shows that the end-to-end BERT-CRF model achieves high F1 score, which reaches 88.56% In this paper, domain high-frequency sub-words are constructed to solve the problem of a large number of unknown words, which reduces the requirements for domain experts and can be easily and cheaply transferred to other fields.The research result of this paper is the practice of information analysis technology in the field of chemistry, which provides important support for the intelligent knowledge retrieval of chemical bond energy.

[Purpose/Significance] The paper aims to study on how different social support and community information in the online Q&A community affecting the cognition and behavior of"lurking"members, and analyze the online experience of lurking users in the online Q&A community so as to improve the activity of the community. [Method/Process] The paper constructed a model of the impact of the matching between questioning information valence and answered social support on platform satisfaction and participation intention of lurking users(third-party observers), used 2 (receiving information valence: positive vs.negative)×2 (The type of social support: affective vs.information)between-group experiments to verify the model. [Result/Conclusion] In the online Q&A community, valence of information and social support have different effects on lurking users' cognition and behavior, and have interactive effects on perceived usefulness and sensitivity.In the positive information, lurking users have higher sensitivity to emotional support than information support, while in the negative information, information support has higher usefulness than emotional support; At the same time, different Q&A matching lead to differences in platform satisfaction and participation willingness of lurking users.

[Purpose/Significance] Exploring the configurations of susceptibility to health misinformation is beneficial.The findings contribute to enhance the ability to cope with health misinformation, and have a positive effect on the health promotion. [Method/Process] Based on Elaboration Likelihood Model and Fogg Behavior Model, this study employed 2×2×2 between-subjects factorial design, and explored the pre-factors by Fuzzy-set Qualitative Comparative Analysis. [Result/Conclusion] The users' ability and motivation are antecedents of the information processing path.High ability and high motivation promote users to process information through the central processing path, while distraction and anxiety promote users to use the peripheral processing path.There are three configuration types of high susceptibility: non-deep thought under strong trigger cues, high information quality, and strong anxiety with strong trigger cues.The antecedent configurations of low susceptibility include two categories: deep thought under weak trigger cues and weak trigger cue mode.In addition, demographic characteristics configuration can also explain susceptibility.From the perspective of information processing, this study explores why users have different susceptibility to health misinformation.The findings can enhance users' immunity to health misinformation and help relevant departments take measures to reduce the impact of health misinformation on individuals and society.

[Purpose/Significance] With the rise of nationwide fitness campaign, the number of sports-injured people is also increasing.Studying the action mechanism of health information avoidance behavior in sports-injured people is conducive to a in-depth understanding of the causes of health information avoidance behavior, so as to provide advice on avoiding the possible harm caused by health information avoidance behavior. [Methods/Process] Based on the research conducted on the influencing factors and behavior patterns of health information avoidance behavior in sports-injured people, this paper collected case data through questionnaires, and further sorted out the relationship between various influencing factors and health information avoidance behavior patterns by using Qualitative Comparative Analysis of Fuzzy Set, so as to explore the configuration of influencing factors of health information avoidance behavior. [Results/Conclusion] According to the different core conditions in each configuration, there are four types of action mechanism of health information avoidance behavior in sports-injured people, which are: "task-driven", "emotion-oriented", "information-susceptible"and"comprehensive-oriented".

[Purpose/Significance] In order to solve the problems of imperfect property legislation and system, lack of effectiveness of scientific data sharing incentive mechanism, and inconsistent management standards in scientific data property management, this study provides a new idea for scientific data property management by constructing a property rights management model for scientific data based on non-fungible tokens(NFTs). [Methods/Processes] First of all, the literature research method was used to analyze the current development status of scientific data property rights management, to summarize the main problems faced in the process of scientific data property rights management, to combine blockchain technology, one of the cutting-edge technologies of current information technology development, to explore the feasibility of NFT application in scientific data property rights management, to build an NFT-based scientific data property rights management model, and to analyze its system processes and related technologies. [Results/Conclusions] The scientific data property rights management system based on NFT technology has the characteristics of high security, perfect incentive mechanism and unified scientific data property rights management standards, so as to better promote the sharing and circulation of scientific data.

[Purpose/Significance] Research tools are indispensable in scientific research activities.From the perspective of tools, mining the characteristics of the application of research tools can observe the changes, status quo and trends of the discipline, which is very important to promote the development of discipline. [Method/Process] This article focused on Python and its software package of library and information science.More than a hundred Python software packages used in the library and information science were divided into 11 categories according to their functions.By using knowledge diffusion theory, network analysis and other analytical methods, this article analyzed the application of Python from three aspects: application profile, application network and application field. [Result/Conclusion] First, the process, stage and motivation of application diffusion are clarified.Second, significant differences in diffusion are revealed at the journal level.Subsequently, the package combination network and package category network are analyzed.Then, the python packages and types with development potential are predicted.Finally, it is proved that the analysis of research tools can objectively judge the development process and trend of the discipline, including the transformation of research paradigm, the change of research methods and technologies, the change of research hotspots, the emergence of new research fields and new research problems, etc.By analyzing the evolution motivation, we can evaluate the development trend of the discipline, clarify the frontier of the discipline and predict the future development direction.

[Purpose/Significance] Open Government data emphasizes on promoting the efficient flow and transformation of government's open data on the basis of multi-subject cooperation.Exploring the data flow and transformation mechanism and practice path of government's open data ecological chain can provide theoretical guidance and experience reference for the flow and transformation of government's open data. [Method/Process] Based on the theory of ecological chain, this paper discussed the concept of government's open data ecological chain, established a data flow and transformation model, and analyzed the practice path of data flow of government's open data ecological chain by taking Shanghai Public Data Open Platform as an example. [Results/Conclusions] The open government data movement promotes the flow and transformation of government's open data resources, and promotes the formation of ecological chain relationship among government's open data subjects.Shanghai takes the public data open platform as the core to cultivate the government's open data ecological chain and to promote the government's open data flow and transformation.It can optimize the design from data governance, policy guarantee, cooperation, technology application to ensure the good operation of the government's open data ecological chain and efficient flow and transformation of government's open data.

[Purpose/Significance] The investigation of measures of countries with advanced practice of socialized use of government open data aims to provide experience and enlightenment for China to promote the socialized use of government data. [Method/Process] This paper investigated the government open data platforms of ten countries via case analysis, summarized six measures to promote the social use of open data, and analyzed experience from three modules of preparation, implementation and effectiveness by referring to the open data barometer. [Result/Conclusion] Finally, the enlightenment of promoting the socialized use of government open data in China is drawn: Construct open data cultural environment to enhance the participation initiative of the whole society; Systematic design of a variety of high-quality measures to improve the breadth and depth of socialized use; Establish a sustainable action mechanism to ensure the effectiveness of the measures.

[Purpose/Significance] With the continuous development of society, the propagation and inheritance mode of intangible cultural heritage have undergone great changes.The short video of intangible cultural heritage based on UGC can display and spread the information of intangible cultural heritage better. [Methods/Process] Taking UGC short videos of intangible cultural heritage as the research object, this paper analyzed the characteristic indicators of the short videos of intangible cultural heritage in the propagation process, and established the measurement model of the propagation force of short videos of intangible cultural heritage based on UGC.In view of the current propagation force of short videos of Peking Opera intangible cultural heritage, this paper also put forward relevant suggestions and countermeasures to improve the propagation force of short videos of intangible cultural heritage. [Result/Conclusions] The propagation force model based on UGC is helpful to understand the current situation of propagation force of short videos of intangible cultural heritage, so as to promote the protection and inheritance of intangible cultural heritage.The magnitude of data obtained in this study is insufficient.The next step is to increase the amount of video data to reduce the results instability caused by the randomness of data.

[Purpose/Significance] The epidemic situation of major infectious diseases is easy to cause public opinion spreading outbreak because of its characteristics of suddenness, sociality and wider spread.Exploring the temporal and spatial differentiation of public opinion in the situation of major infectious diseases, and mastering the differences in the evolution and distribution of people's concerns and emotions in different periods and regions can help the government to identify people's demands, guide social emotions and prevent and control related risks in public opinion of major infectious diseases. [Method/Process] Taking the epidemic situation in COVID-19, Nanjing as an example, the data of public opinion in Weibo was obtained and divided into temporal and spatial regions.Through Top2Vec topic identity and SnowNLP emotion calculation, the temporal and spatial differentiation law of public concern and emotion was described. [Results/Conclusions] In the situation of major infectious diseases, the focus of people's concern shows a trend from pointing to self, pointing to others and then pointing to society in time and space.People's emotional value tends to be positive and increasing in time, and reflects ripple effect in space.People's emotions in marginal areas have certain uncertainty and complexity.

[Purpose/Significance] The increasing popularity and application of short videos have brought convenience to the communication of netizens, but also led to the generation of short video network public opinion, which has brought new challenges to the work of network public opinion guidance and control. [Method/Process] Based on the analysis of relevant subjects and influencing factors of short video network public opinion, Vensim PLE software was used to construct a system dynamics model including three subsystems of netizens, short video platform and government, and combined with"Zhengzhou 7 · 20 Extraordinary rainstorm disaster event", the model was simulated and analyzed, and the process tracking method was used to analyze the causal mechanism of short video network public opinion guidance and control, and to build a theoretical analysis model. [Result/Conclusion] The results show that short video network public opinion is jointly affected by the guidance and control needs, subject participation and coordination mechanism, which can be effectively guided from the aspects of formulating guidance and control plans, strengthening the construction of convergent media matrix, and strengthening effective coordination between subjects.Control short video network public opinion.

[Purpose/Significance] To correctly understand the governance tool attribute of open data of public health emergencies and better realize the sharing and enabling of open data is of the great significance to promote the digital and accurate governance of public health emergencies in China. [Method/Process] First, the privacy leakage risks in public health emergencies data opening and the enabling mechanism of blockchain technology for privacy protection in this context were sorted out; then, privacy protection mode in Public health emergencies data opening empowered by blockchain was constructed; finally, the design of management system was launched to compensate for technical limitations. [Result/Conclusion] The research promotes privacy protection in public health emergency data opening from the dual perspectives of technology and management.

[Purpose/Significance] This paper explores the factors of the survival time of different types of keywords in academic texts from the perspective of semantic function, so as to provide reference for identifying hot topics and technical methods in a field. [Method/Process] First, deep learning was used to identify the term function of keywords.The keywords were divided into problem keywords and method keywords; secondly, a model of influencing factors of keyword survival time was constructed; finally, K-M curve and Cox regression were used to explore the influencing factors of survival time of keywords. [Result/Conclusion] The results show that the number of citations and downloads have positive correlations with the survival time of the problem keywords and method keywords.In addition, journal, funding, and the number of authors and references have significant correlations with the survival time of method keywords, but have no significant correlations with that of problem keywords.

[Purpose/Significance] This paper explores the impact of Altmetric TOP100 list on citations by propensity score matching method to further advance the theoretical and empirical research on Altmetrics. [Method/Process] Based on the balance test and the common support test, this paper used the propensity score matching method to explore the impact of the Altmetric TOP100 list on the citations, and used Mahalanobis matching to test the robustness of the results. [Results/Conclusions] The Altmetric TOP100 list has a significant positive impact on citations; the Altmetric TOP100 list can increase the citations of highly cited papers by 312 on average; the Altmetric TOP100 list can increase the citations of ordinary papers by 134.The Altmetric TOP100 list has a greater impact on the citations of highly cited papers than ordinary papers.